This document contains mathematical formulas and derivations related to probabilistic modeling. It begins by defining the Bernoulli distribution and reparameterizing it using a logit transformation. It then generalizes this to the multinomial distribution and derives properties of the Dirichlet distribution, which is the conjugate prior for the multinomial. The final sections discuss maximum likelihood estimation for these models and properties of the normal-inverse-chi-square distribution, which is the conjugate prior for the normal distribution with unknown mean and variance.